Decomposition of Aryldiazonium Hexafluorophosphates in

Decomposition of Aryldiazonium Hexafluorophosphates in Tetramethylurea. A New Deamination Procedure. Kenneth G. Rutherford, and William A. Redmond...
0 downloads 0 Views 543KB Size
NOTES

568

VOL. 28

changed with those present as chloride ion in solution and for the material to be counted, contained in a glass dish a t a standard every act of substitution 59 of elimination have taken place. distance from the end-window. Immediately before and after Kinetics of the Chloride Ion-promoted Elimination from Chloroeach counting operation the background count was determined tetralone 111.-All runs were carried out by the sealed bulb techand the mean value subtracted from the observed count. The nique using 5.05-ml. aliquots. Titration of acid production was usual counting time was 10 min. and the readings were then corin 30 ml. of acetone previously rendered neutral to lacmoid indirected to number of counts per minute. cator, against a standard solution of sodium methoxide in methRadioactive Count of Pure Tetraethylammonium Radiochloanol. One run was followed by potentiometric titration against ride.-A 1-ml. sample of a 0.0286 M solution of ?rTEt4C136in standard silver nitrate in a titration medium consisting of 30 ml. acetonitrile was counted by the above procedure. The over-all of acetone containing about 1 ml. of 1 N nitric acid. A silver count was found to be 3075 counts per min., which after correction wire electrode and a potassium nitrate-agar bridge to a diptype for the background count of 33 counts per min., gave an actual calomel reference electrode were used. solution count of 3042 per min. Throughout each individual run the first-order rate coefficients Extent of Exchange Accompanying Chloride Ion-promoted Elimwith respect to chlorotetralone I11 fell in value as reaction proination from Bromotetralone 1.-A 50-ml. portion of an acetoceeded. Such a fall was observed and discussed for the bromide nitrile solution 0.0300 M in bromotetralone I and 0.0263 M in ion-promoted elimination from bromotetralone I .z NEt4C136was maintained a t 24.5" for 80 min. A 5-ml. sample was removed and found to require 5.50 ml. of 0.0090 M sodium methoxide for neutralization. The acid formation corresponded to TABLE I 33% elimination. To the remainder of the solution petroleum INITIAL SECOND-ORDER RATE COEFFICIENTS, kz, FOR ACID ether was added and the solution washed well with water until PRODUCTION I N THE REACTIONO F 2-BENZYL-2-CHLORO-4,4the washings were found to be nonradioactive. The petroleum DIMETHYL- TETRALO LONE WITH TETRAETHYLAMMONIUM CHLORIDE ether solution was then evaporated to dryness under reduced presIN SOLVENT ACETONITRILE sure. A solution containing 0.0180 g. of product per ml. of acetonitrile was counted by the standard technique. The actual lO'kz, t, "C. [Chlorotetralone] [NEtaCl] 1. moles-' sec.-l count on the sample was found to be only 6.0 per min. The product was redissolved in petroleum ether and again washed well 61.3 0.0200 0.0769 0.45 with water; a repeat of the above procedure then led to an actual 75.5 ,0100 .00398 2.01Q count of 7.0 per min. showing the small count t o be due to an 75.5 ,0200 ,00961 1.93 actual incorporation of radiochlorine into the organic residue and ,0200 .0769 1.86 75.5 not due to a trace of radiochloride ion impurity. 90.6 .0200 ,00961 8.9 Since the radiochlorine count is extremely low we can to an a Initial second-order rate coefficient for bromide ion production excellent approximation assume the organic residue to be a mix1. 1. moles-' see.-'; kz = A e - E / R T ; A = is 2.06 X ture of bromotetralone I (contaminated with a trace of chloromoles-' see.-'; E = 24.6 kcal./mole. tetralone 111) and endocyclic a,P-unsaturated ketone 11. To 0.0180 g. of residue was added a little acetonitrile and excess piperidine and the mixture allowed to stand for 24 hr. Then 30 An illustrative run is given below; the first-order rate coefml. of acetone and excess nitric acid were added and the bromide ficients, kl (sec.-l), are with respect t o chlorotetralone 111. ion concentration determined by potentiometric titration was equivalent to 3.68 ml. of 0.0100 M silver nitrate. This correTEMPERATURE: 90.6'; 5.05-ML. AIJQUOTS4 T 24'; [CHLOROsponds to a bromotetralone I mole fraction of 0.63 in excellent TETRALONE]: 0.0200 M ; [FEtaCl]: 0.00961; TITERS A ~ R E IN agreement with the 3370 elimination as determined by acid titraML. OF 0.00900 M SAOME tion. Time,min. 0 10 20 30 40 50 60 The concentration of bromotetralone I in the counted solution was 0.0378 M and for a concentration of 0.0286 M the count Titer 0.25 0.76 1.20 1.60 1.97 2.30 2.58 would have been 4.9 per min. The percentage uptake of radiolo5 k l : ,. 7.98 7.58 7.31 7.12 6.90 6.64 chlorine into the bromotetralone is therefore 4.9/3042 X 100%; i . e . , 0.16%. When 33% elimination has occurred the remaining Initial k1 is 8.55 x lO-S~ec.-~,corresponding to an initial secbromotetralone I has 0.16% of its bromine atoms replaced by ond-order rate coefficient of 8.9 X 10-3 1. moles-' see.-'. radiochlorine atoms and for every act of substitution about 200 of elimination have taken place. Acknowledgment.-The authors wish to thank MireExtent of Exchange Accompanying Chloride Ion-promoted ille Coppens for preparation of the 2-benzyl-2-chloro-4,4Elimination from Chlorotetralone II1.-A 60-ml. portion of solution 0.0167 M in chlorotetralone I11 and 0.0182 M in NEtaC13@ dimethyl-1-tetralone. This work was supported in part was maintained at 90.6' for 95 min. A 5-ml. portion was found by grant no. G-14469, National Science Foundation. to require 4.62 ml. of 0.00900M sodium methoxide for neutralization. This corresponds to 50y0elimination reaction having taken place. To the remainder of the solution petroleum ether was added and the solution water washed until the washings were Decomposition of Aryldiazonium Hexafluorononradioactive. The petroleum ether solution was then evapophosphates i n Tetramethylurea. A New rated to dryness under reduced pressure. A solution containing 0.0170 g. of product per ml. of acetoniDeamination Procedure trile was counted by the standard technique. The actual count on the sample was found to be 20.9 per min. To 0.0173 g. of KEXNETH G. RUTHERFORD AND WILLIAM A. REDMOND' sample was added a little acetonitrile and excess piperidine; the solution was maintained a t 90.6" for 24 hr. and then 30 ml. of Department of Chemistry, Essex College, acetone was added and the solution acidified with nitric acid. Assumption University of Windsor, Windsor, Ontario, Canada The chloride ion was found by potentiometric titration to be equivalent to 2.90 ml. of 0.0100 M silver nitrate; this corresponds to a chlorotetralone I11 mole fraction of 0.47, in excellent agreeReceived August 1, 1962 ment with the 50% elimination as determined by acid titration. The concentration of chlorotetralone I11 in the counted solution Interest in this laboratory in the thermal decomposiwas 0.0290 LM and the percentage of C13spresent was 20.9/3042 X tion of aryldiazonium hexafluorophosphates to the 100%; i.e., 0.6976. In determining the amount of exchange allowance must be made for dilution of the C136by hydrogen chlocorresponding aromatic fluorocarbons2 has led to the ride eliminated from the chlorotetralone. At 507, elimination discovery of a new deamination procedure. It has [NEtaC13g]is 0.0182 M and [HCl] is 0.00835 M . The true percentage of chlorine exchange is, therefore, 0.69 [0.0182 0.00835/ (1) Supported by a grant from the National Research Council of Canada. 21 /0.0182; ; . e . , 0.85%. When 50% elimination has occured the ( 2 ) Kenneth G. Rutherford, William Redmond, and James Riganionti. C. S. B., J . Org. Chem., 26, 5149 (1961). remaining chlorotetralone I11 has 0.85% of its chlorine atoms ex-

+

NOT^

FEBRUARY, 1963

569

TABLE I PERCENTYIELDOF DEAMINATIOX PRODUCT"

a

Amine

Ethanol

Aniline o-Toluidine p-Toluidine o-Anisidine Anthranilic acid p-Aminobenzoic acid

5 o-Methylphenetole, ca. 50 o-Methylphenetole, ca. 45 Amount unstated Ethyl benzoate, 53 p-Ethoxybenzoic acid, 50 ethyl benzoate 82 70 87 72-80 68-80

Hypophosphorous acid

60 70-75 77-83

Alkaline formaldehyde

Tetraniethylurea

25-30

60

80 25-30 33 80-85

80 75 25

+

o-Nitroaniline 75-80 p-Nitroaniline 55-60 p-Chloroaniline o-Bromoaniline p-Bromoaniline p- Aminoacetophenone Unless otherwise specified, the amino group has been replaced by hydrogen.

long been known that a primary arom&tic,amino group can be replaced by hydrogen by reduction of its intermediate diazonium c ~ m p o u n d . ~The long established procedures involve reduction of the aromatic diazonium chloride with ethan01,~hypophosphorous acid15or alkaline formaldehyde.6 Sodium stannite,' stannous chloride, and sodium sulfite8have been reported as reducing agents in this reaction, but their use appears to be limited as a general procedure. Recently Meerweing has reported the use of dioxane, dioxolane, diglyme, and tetrahydrofuran as hydride transfer agents in the reduction of aryldiazonium chlorides. The yield of reduction product in each case with the limited number of examples chosen was good. However, the author's attempts to reduce diazonium fluoborates in tertiary amines apparently did not show promise as a useful procedure even though p-chlorobenzenediazonium fluoborate and 2,4,6-trichlorobenzenediazonium fluoborate gave, respectively, and 65% of reduction product. Nevertheless, this is the first case recorded in the literature of reduction of a diazonium salt in nonaqueous media. HendricksonlO has recently reported another nonaqueous reduction in fair yields of a limited number of aromatic diazonium fluoborates with sodium borohydride in methanol. We wish to report the nonaqueous reduction of several aryldiazonium hexafluorophosphates in tetramethylurea. All of the available phosphate salts conveniently decomposed spontaneously and exothermically in this solvent, a phenomenon which was not experienced with several other nonaqueous solvents. Comparisons, where possible, with yields of identical reduction products obtained from the established methods can be observed in the chart. A striking parallel can be observed between tetramethylurea and ethanol reductions. Both methods appear highly desirable as reducing procedures Tvhen electron withdrawing groups are present on the aromatic nucleus of the parent amine. However, in the case of the reduction of ortho- and para-substituted aminobenzoic acids, the tetramethylurea method has a dis(3) Nathan Kornblum, Ow. Reactions, 11, 262 (1944). (4) P.Griess, Phil. Trans. R o y . Soc. London, Ser. A, 154, 683 (1864). ( 5 ) J. Mai, Ber., 36, 163 (1902). (6) R.Q . Brewster and J. A. Poje, J . A m . Cham. Soc.. 61,2418 (1939) (7) P. Friedlander, Ber., 2 2 , 587 (1889). (8) A. Haller, ibid.,18,90 (1885). (9) H.Meerwein. et al., Angew. Chem., 70, 211 (1958). (10) J. B. Hendrickson, J . A m . Chem. Soc.. 83, 1251 (1961).

82 75 62 70 80 72 43

20 10 50

tinct advantage. The yields of reduced product from this method are greater than 80% in the case of both amino acids whereas the aqueous ethanol procedure leads to the formation of benzoate and substituted benzoate ester. As can be seen from Table I, hypophosphorous acid and alkaline formaldehyde procedures lend themselves more readily to the reduction of aromatic primary amines which contain electron donating groups on the ring. Although the reduction of aryldiazonium hexafluorophosphates in tetramethylurea could conceivably follow either an ionic or a free radical path, the former would appear more likely in view of the work of Meerweing and DeTar." Meerwein has obtained very good evidence of hydride transfer in the reaction of 2,4,6-trichlorobenzenediazonium fluoborate with 2phenyl-1,3-dioxolane to give a 5.5% yield of trichlorobenzene and an 80% yield of 2-phenyl-l,3-dioxolenium fluoborate. DeTar has presented strong evidence for a hydride transfer reaction in ethanolic deaminations. The strong parallel of inductive influences between ethanolic and tetramethylurea deaminations would indicate that a similar mechanism is operative. The aryldiazonium hexafluorophosphate presumably breaks down in tetramethylurea ionically according to equation l. ArNzPFB+Are

+ X2 + PFse

(1)

The aryl cation, depending on its reactivity can either abstract hydride from the solvent (equation 2), couple with it to form a salt (equation 3), or react with PFse to yield the fluoroaromatic and PFs (equation 4). With the highly reactive aryl cations, the activation energy requirement for hydride transfer would be met more readily. The resultant urea carbonium ion could conceivably react with the PFee ion to yield PFb and fluorinated tetramethylurea, although no attempt has been made thus far to isolate and characterize such a compound. The less reactive cations would not be sufficiently energetic for predominant hydride transfer and competing reactions 3 and 4 become more likely. No evidence of reaction 4 has been observed with any of the parent amines studied, although this does not preclude the possibility of traces of fluoroaromatic present. Evidence for reaction 3, however, is concrete ( 1 1 ) D. F. DeTar and T . Kosuge, ibid.. 80, 6072 (1958)

NOTES

570 HIC

+

Ar@

0

CHs

‘N-LN’

/

Hac

CHs ‘WH2 ArH

+

0

CH,

N ’& -N -\

/

\

CHs

0

HaC

AP

+

I1

CHs

CHs

/ N--CN / \ \

an electron transfer from the PFse ion to the positively charged N1: species. This would leave phosphorus pentafluoride and a fluorine atom. Regardless of the fate of the fluorine atom, reformation of the hexafluorophosphate anion would be very unlikely. Again, an electron transfer from the solvent to the Nz@ species would render the solvent positively charged. This would have a stronger tendency to react with the PFP ion than the aromatic free radical, again producing PFs. Similar salts as well as reduction products have been isolated from the decomposition of the diazonium hexafluorophosphates derived from p-nitroaniline, p-toluidine, and aniline. It is interesting to note that Meenveinla decomposed the diazonium fluoborates derived from pnitroaniline and pchloroaniline in tetramethylthiourea. No reduction product was reported-instead there were obtained in high yields tetramethylthiouronium fluoborate salts of the proposed structure:

+

\

--+

CHa

H3C

CH;

X = N02,Cl

H3C

Ar-0-C

Evidently the much higher nucleophilic property of the sulfur atom over oxygen as well as its greater size makes it more susceptible to attack by the aryl cation and hence a reaction analogous to equation 3 becomes predominant. Thus tetramethylthiourea is not a hydride transfer agent even with the more active aryl cations.

CHa

/ Q

NmPFe,*

/ \

CHI

CH3

in the case of at least one amine studied in detail thus far. The diazonium hexafluorophosphate derived from o-toluidine gives on decomposition in tetramethylurea a 6lY0 yield of o-methylphenoxytetramethylamidinium hexafluorophosphate I. The structure of this salt was

r i

proved by elemental analysis, and vigorous base hydrolysis to o-cresol, and dimethylamine. The PFse ion was determined quantitatively by tetraphenylarsonium chloride reagent.la An amidinium structure is proposed since the infrared spectrogram showed a strong absorption a t 5.90 p . The presence of PF,e ion in this salt renders a free radical reaction highly unlikely. A free radical breakdown of the aryldiazonium hexafluorophosphate would occur according to equation 5 . ArN2PFs -+ Ar.

VOL.28

+ :N&. + I’Fee

(5)

To explain the evolution of nitrogen (97% of the theoretical amount), one must envisage a t least in part (12) The PFP ion is determined quantitatively as tetraphenylarsonium hexafluorophosphate.

Experimental” Decomposition of Aryldiazonium Hexafluorophosphates.The aryldiazonium hexafluorophosphates were prepared as previously reported.2 The dry salts were decomposed by the portionwise addition of 50 g. to a beaker which contained tetramethylurea (200 ml.) at room temperature. Usually a rapid evolution of nitrogen was noted and the temperature of the reaction mixture rose quite rapidly. In cases where the reaction was only moderately exothermic, the mixture was warmed t o increase the rate. The temperature was not allowed to rise above 65’ in any of the runs. When all sign of reaction had ceased as evidenced by the lack of nitrogen evolution (about 1.5 hr.), the solution was poured into 1 1. of water. The reduction product in most cases was recovered by distillation with steam followed by separation from the water layer and fractionation. In the case of the reduction of the amino benzoic acids, the reaction mixture was poured on a mixture of cracked ice and hydrochloric acid. The benzoic acid was recovered by filtration. o-Methylphenoxytetramethylamidinium Hexafluorophosphate (I).+-Toluidinediazonium hexafluorophosphate was decomposed in the manner described above. The reaction mixture was poured on a mixture of ice and hydrochloric acid. The precipitate obtained was recovered by filtration and air dried. The red color was removed by treatment of a chloroform solution of the compound with Norite. Recrystallization from a chloroformpetroleum ether (b.p. 30-60’) solvent pair gave an average of 61% yield on several runs of colorless o-methylphenoxytetramethylamidinium hexafluorophosphate, m.p. 144-145’. Anal. Calculated for CLZHIQONZPF~: C, 40.90; H, 5.44; N, 7.95; PF8, 41.16. Found: C, 40.57; H, 5.24; N, 7.75; PFs, 41.04. Hydrolysis of 1.-The hydrolysis of I (15 g., 0.043 mole) was accomplished in 100 ml. of 4 N sodium hydroside in a 250-ml. (13) H. Meerwein, et ul., Ber.. DO, 851 (1957). (14) All melting points are uncorrected.

NOTES

FEBRUARY, 1963 round-bottomed flask fitted with a condenser. A rubber connecting tube was led from the top of the condenser into a gas washing bottle which was fitted with a sintered glass disk. To the washing bottle was added 150 ml. of anhydrous ether. While heating the reaction flask a t reflux temperature, a gas was evolved which dissolved in the ether. After 4 hr. the apparatus was disassembled and dry hydrogen chloride gas was bubbled through the ethereal solution. The precipitate which formed was isolated by rapid filtration and dried in vacuo. A yield of 3.2 g. (48.4%) of dimethylamine hydrochloride, m.p. 168-171 ', was obtained. Addition of Dry Ice to the reaction mixture resulted in the formation of a pink organic layer which when separated yielded 7.5 g. (83%) of o-cresol. The p-bromobenzenesulfonate derivative melted a t 78". A mixture melting point determination with an authentic sample showed no depression. The dibromo derivative (m.p. 56') likewise showed no depression in melting point when mixed with an equal portion of authentic material.

Acknowledgment-The authors wish to thank the Ozark-Mahoning Co. of Tulsa, Oklahoma, for supplies of hexafluorophosphoric acid and the Grand River Chemical Division of John Deere and Co., Pryor, Oklahoma, for supplies of tetramethylurea.

The Solvolysis of Stigmasteryl Tosylate JOHNA. STEELEAND ERICHMOSETTIG' National Institute of Arthritis and Metabolic Diseases, National Institutes of Health, Bethesda 14, Maryland Received August 9, 1962

In connection with studies on the identification of an unknown house-fly sterollZaa sample of high purity psitosterol was required. All commercial samples of the sterol showed, by gas chromatography, a major impurityzb varying, in content, from 10-30%,. We therefore converted stigmasterol to 0-sitosterol according to published procedures3 but, again, could not obtain a preparation of satisfactory purity. &,5

RO

-

&(9

OR

I n repeating the solvolysis of stigmasteryl tosylate (I) in methanol and chromatography on alumina,3b the product obtained could be resolved further by rechromatography on Florisil into i-stigmasteryl methyl ether (11) and, as minor components, stigmasteryl methyl ether (III), stigmasterol and a hydrocarbon (IV). Riegel and co-workers4 described a similar hy(1) Deceased, May 31,1962. (2) (a) Dr. William Robbins and staff, Insect Physiology Laboratory, ARS, U. S. Department of Agriculture, Beltsville, Md.; (b) The nature of this impurity is the subject of a forthcoming publication by M. J. Thompson, 5. J. Louloudes, W. Robbins, J. A. Waters, J. A . Steele, and E. Mosettig. (3)(a) S. Bernstein and E. S. Wallis, J . O w . Chem., 2, 341 (1938); (b) E. Fernholz and W. L. Ruigh, J . Am. Chem. Soc., 62, 3346 (1940); ( 0 ) B. Riegel, E. W. Meyer, and J. Beiswanger, ibid., 66, 325 (1943). (4) B. Riegel, G. P. Hager, and B. L. Zenity, ibid., 68, 2582 (1946).

571

drocarbon in the i-cholesterol series to which they assigned the A%cholestene structure on the basis of its physical and chemical behavior. The structure 3,5cyclo-6,22-sitostadiene has been assigned to hydrocarbon IV by analogy.6 Since chromatography of the reaction mixture on Florisil directly gave only i-stigmasteryl methyl ether, stigmasteryl methyl ether, and stigmasterol, it would appear that the hydrocarbon IX is an artifact resulting from contact of I1 with an alumina column. Hydrogenation of i-stigmasteryl methyl ether followed by rearrangement with zinc acetate in boiling acetic acid gave, after one recrystallization, 22,234hydrostigmasteryl acetate. Hydrolysis of acetate IX gave pure 22,23-dihydrostigmasterol. The solvolysis of stigmasteryl tosylate in acetonewater was carried out with slight modification of the procedure described for the preparation of i-ergosterol and i-dehydroergosterol.6 Chromatography of the mixture on Florisil gave i-stigmasterol V and stigmgsterol VI. Hydrogenation of i-stigmasterol V afforded i22,23-dihydrostigmasterol (i-@-sitosterol) VIII. Rearrangement of this product with zinc acetate in boiling acetic acid gave 22,23-dihydrostigmasteryl acetate IX. The over-all yield of 22,23-dihydrostigmasterol obtained via the i-sterol (70%) is better than that obtained via the i-stigmasteryl methyl ether (63%). In addition, the route via the i-sterol permits the recovery of reusable stigmasterol. The 22,23-dihydrostigmasterol obtained through either procedure gave only one peak in gas-liquid chromatography which differed from that of stigmasterol. Although there have been many examples of @-sitosterol isolated from plant sources, the purity of these materials may be open to question. The conversion of stigmasterol to 0-sitosterol via the i-sterol is a convenient procedure, affording good yields and leading to a product of very high purity. Experimental? Stigmasteryl Tosylate (I).-To a solution of 15.0 g. of dry stigmasterol in 200 ml. of dry pyridine was added 18.0 g. of p-toluenesulfonyl chloride [freshly recrystallized from petroleum ether (b .p. 60-70'). The reaction mixture was allowed to stand overnight in the dark a t room temperature and poured into 1 1. of ice-cold 5% aqueous potassium bicarbonate solution. The resulting solid tosylate was collected by filtration, washed with water, and dried in a vacuum oven a t 60". Recrystallization from dry acetone gave (I)as colorless crystals, m.p. 147-148", [ a ] -~49'. i-Stigmasteryl Methyl Ether (II).-To a refluxing solution of 8.0 g. of fused potassium acetate in 400 ml. of dry methanol was added 8.0 g. of finely pulverized stigmsnteryl tosylate. The mixture was refluxed for 3 hr. After removal of the solvent in vacuo, the residue was extracted with ether, washed with water, 5% aqueous potassium bicarbonate, water, dried over potassium carbonate, and evaporated in vacuo leaving 6.0 g. of oily material. This oily residue was dissolved in petroleum ether, adsorbed on a ( 5 ) The ultraviolet spectrum of IX showed only end absorption, thus ruling out a 3,5-diene structure. (6) W. R. Nes and J. A. Steele, J . Org. Chem., 22, 1457 (1957). (7) All melting points were determined on a Kofler block and are uncorrected. Analyses are b y the Analytical Service Laboratory of this Institute under the direction of Mr. Harold G. McCann. Rotations were taken in 1% chloroform solutions s t 20° b y Mre. E. Peake. The infrared spectra were determined on a Perkin-Elmer double beam spectrophotometer (Model 21) in carbon disulfide b y Mr. H. K. Miller with the assistance of hfrs. C. I. Wright. The gas chromatographic analyses were carried out on a BarberColman gas chromatograph with SE-30 support. "Woelm" neutral alumina, activity grade I. and Florisil (60-100 mesh) purchased from the Floridin Co., Tallahassee. Fla. were used for chromatography. Solvents were purified, distilled, and dried with the help of Mr. J. Lyons, with the exception of ethanol and chloroform which were the usual commercial grades of reagent material.